The homeodomain family of transcription factors plays a fundamental role in a diverse set of functions that include body plan specification, pattern formation, and cell fate determination during metazoan development. Members of this family are characterized by a helix-turn-helix DNA-binding motif known as the homeodomain. Homeodomain proteins regulate various cellular processes by specifically binding to the transcriptional control region of a target gene. These proteins have been conserved across a diverse range of species, from yeast to human. A number of inherited human disorders are caused by mutations in homeodomain-containing proteins. For example, mutations in a number of forkhead transcription factors are associated with the development of inherited diseases in humans. Two closely related genes, FOXP2 and FOXP3, are implicated in two completely different human disorders. A point mutation in the forkhead domain of FOXP2 (R553H) is responsible for a severe speech and language disorder, while a series of missense mutations distributed over the forkhead domain of FOXP3 cause a fatal disorder called IPEX, characterized by immune dysregulation, polyendocrinopathy, and enteropathy. Homology model building techniques were used to generate atomic structures of FOXP2 and FOXP3, using the solution structures of the forkhead domain of the adipocyte-transcription factor FREAC-11 and AFX as templates. The impact of these disease-causing missense mutations on the three-dimensional structure, stability, and surface electrostatic charge distribution of the forkhead domains were examined. The missense mutations R553H in FOXP2 and R397W in FOXP3 dramatically alter the electrostatic potentials of the molecular surface of their respective forkhead domains. Similar studies are currently under way to examine the effect of mutations in the FOXH1 gene on the atomic structures of the resulting protein. The FOXH1 studies are being performed in collaboration with Dr. Max Muenke, Medical Genetics Branch, NHGRI, as part of a broader study aimed at understanding the function of this protein in development. Finally, the workgroup continues to oversee the curation of the Homeodomain resource, a searchable collection of information for the homeodomain protein family. The resource is organized in a compact form and provides user-friendly interfaces for both querying the component databases and assembling customized datasets. A new release will be made available in the fall of 2005 that contains 4,385 full-length homeodomain-containing sequences covering 546 distinct organisms, 50 experimentally-derived structures, 90 homeodomain interactions, 91 homeodomain DNA-binding sites, 181 homeodomain proteins implicated in human genetic disorders, and 38 homeodomain proteins with documented allelic variants. The information available on both allelic variants and experimentally-derived protein-protein interactions has been significantly expanded, encompassing information from new sources, such as the Human Gene Mutation Database and the Biomolecular Interaction Network (BIND). All experimentally-derived protein-protein interaction data derived from the literature has been contributed to BIND as well. An extensive rewrite of the underlying code was undertaken, porting the database to the Oracle platform, which facilitated the ability to process complex searches and increase the efficiency of searches, in terms of speed. The Homeodomain Resource is freely available through the World Wide Web at http://research.nhgri.nih.gov/homeodomain/.